In [8]:
structure = cfgs.MODEL.STRUCTURE[0]
content_model = MultiOutputModel(*cfgs.IMAGE.RESOLUTION, structure)
content_model.build_model("InceptionV3")
keras.utils.plot_model(content_model.model, show_shapes=True)
Out[8]:
from cfgs.content_classification_v2 import get_cfg
cfgs = get_cfg()
print('\n', cfgs, '\n')
CLASSES: [{'Genre': {0: 'Comedy', 1: 'Action', 2: 'Documentary', 3: 'Crime', 4: 'Animation', 5: 'Horror'}, 'Rating': {0: 'value'}, 'Year': {0: 'value'}}]
HYPERPARAMS:
BATCH_SIZE_TEST: 64
BATCH_SIZE_TR: 64
BATCH_SIZE_VAL: 64
EPOCHS: 30
LR: 0.0001
TEST_SPLIT: 0.9
VAL_SPLIT: 0.9
IMAGE:
RESOLUTION: (150, 200, 3)
MODEL:
BACKBONE: InceptionV3
STRUCTURE: [{0: {'type': 'Genre', 'outNeurons': 6, 'outActivation': 'sigmoid', 'loss': 'categorical_crossentropy', 'weight': 1, 'metric': 'accuracy'}, 1: {'type': 'Rating', 'outNeurons': 1, 'outActivation': 'linear', 'loss': 'mse', 'weight': 1, 'metric': 'mae'}, 2: {'type': 'Year', 'outNeurons': 1, 'outActivation': 'linear', 'loss': 'mse', 'weight': 1, 'metric': 'mae'}}]
dataset_dict = cfgs.CLASSES[0]
dataset_dict['Genre_alias'] = dict((g, i) for i, g in dataset_dict['Genre'].items())
import pandas as pd
from data_processing.data_generator import DataGenerator
from sklearn.preprocessing import StandardScaler
df = pd.read_pickle("/home/robotics/content-classification/data_processing/balanced_data.pkl")
data_generator = DataGenerator(df, dataset_dict, cfgs.HYPERPARAMS.TEST_SPLIT, cfgs.HYPERPARAMS.VAL_SPLIT, cfgs.IMAGE.RESOLUTION)
data_generator.df.head()
| Index | Name | Year | Genre | Rating | Poster_path | |
|---|---|---|---|---|---|---|
| 0 | tt0077838 | The Last Waltz | 1978 | 2 | 8.2 | /home/robotics/Documents/data/Posters/1978/tt0... |
| 1 | tt0218080 | Agent Red | 2000 | 1 | 3.5 | /home/robotics/Documents/data/Posters/2000/tt0... |
| 2 | tt0378889 | Tom, Tom, the Piper's Son | 1969 | 0 | 6.2 | /home/robotics/Documents/data/Posters/1969/tt0... |
| 3 | tt1183696 | Second Skin | 2009 | 2 | 6.6 | /home/robotics/Documents/data/Posters/2009/tt1... |
| 4 | tt0068549 | Essene | 1972 | 2 | 7.1 | /home/robotics/Documents/data/Posters/1972/tt0... |
train_idx, valid_idx, test_idx = data_generator.split_dataset()
from tensorflow.keras.optimizers import Adam
from tensorflow import keras
from model.architecture import MultiOutputModel
structure = cfgs.MODEL.STRUCTURE[0]
content_model = MultiOutputModel(*cfgs.IMAGE.RESOLUTION, structure)
content_model.build_model("InceptionV3")
keras.utils.plot_model(content_model.model, show_shapes=True)
# opt = SGD()
opt = Adam(lr=cfgs.HYPERPARAMS.LR, decay=cfgs.HYPERPARAMS.LR/cfgs.HYPERPARAMS.EPOCHS)
content_model.compile_model(opt)
content_model.model.summary()
Model: "content_classification"
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
input_1 (InputLayer) [(None, 150, 200, 3) 0
__________________________________________________________________________________________________
conv2d (Conv2D) (None, 74, 99, 32) 864 input_1[0][0]
__________________________________________________________________________________________________
batch_normalization (BatchNorma (None, 74, 99, 32) 96 conv2d[0][0]
__________________________________________________________________________________________________
activation (Activation) (None, 74, 99, 32) 0 batch_normalization[0][0]
__________________________________________________________________________________________________
conv2d_1 (Conv2D) (None, 72, 97, 32) 9216 activation[0][0]
__________________________________________________________________________________________________
batch_normalization_1 (BatchNor (None, 72, 97, 32) 96 conv2d_1[0][0]
__________________________________________________________________________________________________
activation_1 (Activation) (None, 72, 97, 32) 0 batch_normalization_1[0][0]
__________________________________________________________________________________________________
conv2d_2 (Conv2D) (None, 72, 97, 64) 18432 activation_1[0][0]
__________________________________________________________________________________________________
batch_normalization_2 (BatchNor (None, 72, 97, 64) 192 conv2d_2[0][0]
__________________________________________________________________________________________________
activation_2 (Activation) (None, 72, 97, 64) 0 batch_normalization_2[0][0]
__________________________________________________________________________________________________
max_pooling2d (MaxPooling2D) (None, 35, 48, 64) 0 activation_2[0][0]
__________________________________________________________________________________________________
conv2d_3 (Conv2D) (None, 35, 48, 80) 5120 max_pooling2d[0][0]
__________________________________________________________________________________________________
batch_normalization_3 (BatchNor (None, 35, 48, 80) 240 conv2d_3[0][0]
__________________________________________________________________________________________________
activation_3 (Activation) (None, 35, 48, 80) 0 batch_normalization_3[0][0]
__________________________________________________________________________________________________
conv2d_4 (Conv2D) (None, 33, 46, 192) 138240 activation_3[0][0]
__________________________________________________________________________________________________
batch_normalization_4 (BatchNor (None, 33, 46, 192) 576 conv2d_4[0][0]
__________________________________________________________________________________________________
activation_4 (Activation) (None, 33, 46, 192) 0 batch_normalization_4[0][0]
__________________________________________________________________________________________________
max_pooling2d_1 (MaxPooling2D) (None, 16, 22, 192) 0 activation_4[0][0]
__________________________________________________________________________________________________
conv2d_8 (Conv2D) (None, 16, 22, 64) 12288 max_pooling2d_1[0][0]
__________________________________________________________________________________________________
batch_normalization_8 (BatchNor (None, 16, 22, 64) 192 conv2d_8[0][0]
__________________________________________________________________________________________________
activation_8 (Activation) (None, 16, 22, 64) 0 batch_normalization_8[0][0]
__________________________________________________________________________________________________
conv2d_6 (Conv2D) (None, 16, 22, 48) 9216 max_pooling2d_1[0][0]
__________________________________________________________________________________________________
conv2d_9 (Conv2D) (None, 16, 22, 96) 55296 activation_8[0][0]
__________________________________________________________________________________________________
batch_normalization_6 (BatchNor (None, 16, 22, 48) 144 conv2d_6[0][0]
__________________________________________________________________________________________________
batch_normalization_9 (BatchNor (None, 16, 22, 96) 288 conv2d_9[0][0]
__________________________________________________________________________________________________
activation_6 (Activation) (None, 16, 22, 48) 0 batch_normalization_6[0][0]
__________________________________________________________________________________________________
activation_9 (Activation) (None, 16, 22, 96) 0 batch_normalization_9[0][0]
__________________________________________________________________________________________________
average_pooling2d (AveragePooli (None, 16, 22, 192) 0 max_pooling2d_1[0][0]
__________________________________________________________________________________________________
conv2d_5 (Conv2D) (None, 16, 22, 64) 12288 max_pooling2d_1[0][0]
__________________________________________________________________________________________________
conv2d_7 (Conv2D) (None, 16, 22, 64) 76800 activation_6[0][0]
__________________________________________________________________________________________________
conv2d_10 (Conv2D) (None, 16, 22, 96) 82944 activation_9[0][0]
__________________________________________________________________________________________________
conv2d_11 (Conv2D) (None, 16, 22, 32) 6144 average_pooling2d[0][0]
__________________________________________________________________________________________________
batch_normalization_5 (BatchNor (None, 16, 22, 64) 192 conv2d_5[0][0]
__________________________________________________________________________________________________
batch_normalization_7 (BatchNor (None, 16, 22, 64) 192 conv2d_7[0][0]
__________________________________________________________________________________________________
batch_normalization_10 (BatchNo (None, 16, 22, 96) 288 conv2d_10[0][0]
__________________________________________________________________________________________________
batch_normalization_11 (BatchNo (None, 16, 22, 32) 96 conv2d_11[0][0]
__________________________________________________________________________________________________
activation_5 (Activation) (None, 16, 22, 64) 0 batch_normalization_5[0][0]
__________________________________________________________________________________________________
activation_7 (Activation) (None, 16, 22, 64) 0 batch_normalization_7[0][0]
__________________________________________________________________________________________________
activation_10 (Activation) (None, 16, 22, 96) 0 batch_normalization_10[0][0]
__________________________________________________________________________________________________
activation_11 (Activation) (None, 16, 22, 32) 0 batch_normalization_11[0][0]
__________________________________________________________________________________________________
mixed0 (Concatenate) (None, 16, 22, 256) 0 activation_5[0][0]
activation_7[0][0]
activation_10[0][0]
activation_11[0][0]
__________________________________________________________________________________________________
conv2d_15 (Conv2D) (None, 16, 22, 64) 16384 mixed0[0][0]
__________________________________________________________________________________________________
batch_normalization_15 (BatchNo (None, 16, 22, 64) 192 conv2d_15[0][0]
__________________________________________________________________________________________________
activation_15 (Activation) (None, 16, 22, 64) 0 batch_normalization_15[0][0]
__________________________________________________________________________________________________
conv2d_13 (Conv2D) (None, 16, 22, 48) 12288 mixed0[0][0]
__________________________________________________________________________________________________
conv2d_16 (Conv2D) (None, 16, 22, 96) 55296 activation_15[0][0]
__________________________________________________________________________________________________
batch_normalization_13 (BatchNo (None, 16, 22, 48) 144 conv2d_13[0][0]
__________________________________________________________________________________________________
batch_normalization_16 (BatchNo (None, 16, 22, 96) 288 conv2d_16[0][0]
__________________________________________________________________________________________________
activation_13 (Activation) (None, 16, 22, 48) 0 batch_normalization_13[0][0]
__________________________________________________________________________________________________
activation_16 (Activation) (None, 16, 22, 96) 0 batch_normalization_16[0][0]
__________________________________________________________________________________________________
average_pooling2d_1 (AveragePoo (None, 16, 22, 256) 0 mixed0[0][0]
__________________________________________________________________________________________________
conv2d_12 (Conv2D) (None, 16, 22, 64) 16384 mixed0[0][0]
__________________________________________________________________________________________________
conv2d_14 (Conv2D) (None, 16, 22, 64) 76800 activation_13[0][0]
__________________________________________________________________________________________________
conv2d_17 (Conv2D) (None, 16, 22, 96) 82944 activation_16[0][0]
__________________________________________________________________________________________________
conv2d_18 (Conv2D) (None, 16, 22, 64) 16384 average_pooling2d_1[0][0]
__________________________________________________________________________________________________
batch_normalization_12 (BatchNo (None, 16, 22, 64) 192 conv2d_12[0][0]
__________________________________________________________________________________________________
batch_normalization_14 (BatchNo (None, 16, 22, 64) 192 conv2d_14[0][0]
__________________________________________________________________________________________________
batch_normalization_17 (BatchNo (None, 16, 22, 96) 288 conv2d_17[0][0]
__________________________________________________________________________________________________
batch_normalization_18 (BatchNo (None, 16, 22, 64) 192 conv2d_18[0][0]
__________________________________________________________________________________________________
activation_12 (Activation) (None, 16, 22, 64) 0 batch_normalization_12[0][0]
__________________________________________________________________________________________________
activation_14 (Activation) (None, 16, 22, 64) 0 batch_normalization_14[0][0]
__________________________________________________________________________________________________
activation_17 (Activation) (None, 16, 22, 96) 0 batch_normalization_17[0][0]
__________________________________________________________________________________________________
activation_18 (Activation) (None, 16, 22, 64) 0 batch_normalization_18[0][0]
__________________________________________________________________________________________________
mixed1 (Concatenate) (None, 16, 22, 288) 0 activation_12[0][0]
activation_14[0][0]
activation_17[0][0]
activation_18[0][0]
__________________________________________________________________________________________________
conv2d_22 (Conv2D) (None, 16, 22, 64) 18432 mixed1[0][0]
__________________________________________________________________________________________________
batch_normalization_22 (BatchNo (None, 16, 22, 64) 192 conv2d_22[0][0]
__________________________________________________________________________________________________
activation_22 (Activation) (None, 16, 22, 64) 0 batch_normalization_22[0][0]
__________________________________________________________________________________________________
conv2d_20 (Conv2D) (None, 16, 22, 48) 13824 mixed1[0][0]
__________________________________________________________________________________________________
conv2d_23 (Conv2D) (None, 16, 22, 96) 55296 activation_22[0][0]
__________________________________________________________________________________________________
batch_normalization_20 (BatchNo (None, 16, 22, 48) 144 conv2d_20[0][0]
__________________________________________________________________________________________________
batch_normalization_23 (BatchNo (None, 16, 22, 96) 288 conv2d_23[0][0]
__________________________________________________________________________________________________
activation_20 (Activation) (None, 16, 22, 48) 0 batch_normalization_20[0][0]
__________________________________________________________________________________________________
activation_23 (Activation) (None, 16, 22, 96) 0 batch_normalization_23[0][0]
__________________________________________________________________________________________________
average_pooling2d_2 (AveragePoo (None, 16, 22, 288) 0 mixed1[0][0]
__________________________________________________________________________________________________
conv2d_19 (Conv2D) (None, 16, 22, 64) 18432 mixed1[0][0]
__________________________________________________________________________________________________
conv2d_21 (Conv2D) (None, 16, 22, 64) 76800 activation_20[0][0]
__________________________________________________________________________________________________
conv2d_24 (Conv2D) (None, 16, 22, 96) 82944 activation_23[0][0]
__________________________________________________________________________________________________
conv2d_25 (Conv2D) (None, 16, 22, 64) 18432 average_pooling2d_2[0][0]
__________________________________________________________________________________________________
batch_normalization_19 (BatchNo (None, 16, 22, 64) 192 conv2d_19[0][0]
__________________________________________________________________________________________________
batch_normalization_21 (BatchNo (None, 16, 22, 64) 192 conv2d_21[0][0]
__________________________________________________________________________________________________
batch_normalization_24 (BatchNo (None, 16, 22, 96) 288 conv2d_24[0][0]
__________________________________________________________________________________________________
batch_normalization_25 (BatchNo (None, 16, 22, 64) 192 conv2d_25[0][0]
__________________________________________________________________________________________________
activation_19 (Activation) (None, 16, 22, 64) 0 batch_normalization_19[0][0]
__________________________________________________________________________________________________
activation_21 (Activation) (None, 16, 22, 64) 0 batch_normalization_21[0][0]
__________________________________________________________________________________________________
activation_24 (Activation) (None, 16, 22, 96) 0 batch_normalization_24[0][0]
__________________________________________________________________________________________________
activation_25 (Activation) (None, 16, 22, 64) 0 batch_normalization_25[0][0]
__________________________________________________________________________________________________
mixed2 (Concatenate) (None, 16, 22, 288) 0 activation_19[0][0]
activation_21[0][0]
activation_24[0][0]
activation_25[0][0]
__________________________________________________________________________________________________
conv2d_27 (Conv2D) (None, 16, 22, 64) 18432 mixed2[0][0]
__________________________________________________________________________________________________
batch_normalization_27 (BatchNo (None, 16, 22, 64) 192 conv2d_27[0][0]
__________________________________________________________________________________________________
activation_27 (Activation) (None, 16, 22, 64) 0 batch_normalization_27[0][0]
__________________________________________________________________________________________________
conv2d_28 (Conv2D) (None, 16, 22, 96) 55296 activation_27[0][0]
__________________________________________________________________________________________________
batch_normalization_28 (BatchNo (None, 16, 22, 96) 288 conv2d_28[0][0]
__________________________________________________________________________________________________
activation_28 (Activation) (None, 16, 22, 96) 0 batch_normalization_28[0][0]
__________________________________________________________________________________________________
conv2d_26 (Conv2D) (None, 7, 10, 384) 995328 mixed2[0][0]
__________________________________________________________________________________________________
conv2d_29 (Conv2D) (None, 7, 10, 96) 82944 activation_28[0][0]
__________________________________________________________________________________________________
batch_normalization_26 (BatchNo (None, 7, 10, 384) 1152 conv2d_26[0][0]
__________________________________________________________________________________________________
batch_normalization_29 (BatchNo (None, 7, 10, 96) 288 conv2d_29[0][0]
__________________________________________________________________________________________________
activation_26 (Activation) (None, 7, 10, 384) 0 batch_normalization_26[0][0]
__________________________________________________________________________________________________
activation_29 (Activation) (None, 7, 10, 96) 0 batch_normalization_29[0][0]
__________________________________________________________________________________________________
max_pooling2d_2 (MaxPooling2D) (None, 7, 10, 288) 0 mixed2[0][0]
__________________________________________________________________________________________________
mixed3 (Concatenate) (None, 7, 10, 768) 0 activation_26[0][0]
activation_29[0][0]
max_pooling2d_2[0][0]
__________________________________________________________________________________________________
conv2d_34 (Conv2D) (None, 7, 10, 128) 98304 mixed3[0][0]
__________________________________________________________________________________________________
batch_normalization_34 (BatchNo (None, 7, 10, 128) 384 conv2d_34[0][0]
__________________________________________________________________________________________________
activation_34 (Activation) (None, 7, 10, 128) 0 batch_normalization_34[0][0]
__________________________________________________________________________________________________
conv2d_35 (Conv2D) (None, 7, 10, 128) 114688 activation_34[0][0]
__________________________________________________________________________________________________
batch_normalization_35 (BatchNo (None, 7, 10, 128) 384 conv2d_35[0][0]
__________________________________________________________________________________________________
activation_35 (Activation) (None, 7, 10, 128) 0 batch_normalization_35[0][0]
__________________________________________________________________________________________________
conv2d_31 (Conv2D) (None, 7, 10, 128) 98304 mixed3[0][0]
__________________________________________________________________________________________________
conv2d_36 (Conv2D) (None, 7, 10, 128) 114688 activation_35[0][0]
__________________________________________________________________________________________________
batch_normalization_31 (BatchNo (None, 7, 10, 128) 384 conv2d_31[0][0]
__________________________________________________________________________________________________
batch_normalization_36 (BatchNo (None, 7, 10, 128) 384 conv2d_36[0][0]
__________________________________________________________________________________________________
activation_31 (Activation) (None, 7, 10, 128) 0 batch_normalization_31[0][0]
__________________________________________________________________________________________________
activation_36 (Activation) (None, 7, 10, 128) 0 batch_normalization_36[0][0]
__________________________________________________________________________________________________
conv2d_32 (Conv2D) (None, 7, 10, 128) 114688 activation_31[0][0]
__________________________________________________________________________________________________
conv2d_37 (Conv2D) (None, 7, 10, 128) 114688 activation_36[0][0]
__________________________________________________________________________________________________
batch_normalization_32 (BatchNo (None, 7, 10, 128) 384 conv2d_32[0][0]
__________________________________________________________________________________________________
batch_normalization_37 (BatchNo (None, 7, 10, 128) 384 conv2d_37[0][0]
__________________________________________________________________________________________________
activation_32 (Activation) (None, 7, 10, 128) 0 batch_normalization_32[0][0]
__________________________________________________________________________________________________
activation_37 (Activation) (None, 7, 10, 128) 0 batch_normalization_37[0][0]
__________________________________________________________________________________________________
average_pooling2d_3 (AveragePoo (None, 7, 10, 768) 0 mixed3[0][0]
__________________________________________________________________________________________________
conv2d_30 (Conv2D) (None, 7, 10, 192) 147456 mixed3[0][0]
__________________________________________________________________________________________________
conv2d_33 (Conv2D) (None, 7, 10, 192) 172032 activation_32[0][0]
__________________________________________________________________________________________________
conv2d_38 (Conv2D) (None, 7, 10, 192) 172032 activation_37[0][0]
__________________________________________________________________________________________________
conv2d_39 (Conv2D) (None, 7, 10, 192) 147456 average_pooling2d_3[0][0]
__________________________________________________________________________________________________
batch_normalization_30 (BatchNo (None, 7, 10, 192) 576 conv2d_30[0][0]
__________________________________________________________________________________________________
batch_normalization_33 (BatchNo (None, 7, 10, 192) 576 conv2d_33[0][0]
__________________________________________________________________________________________________
batch_normalization_38 (BatchNo (None, 7, 10, 192) 576 conv2d_38[0][0]
__________________________________________________________________________________________________
batch_normalization_39 (BatchNo (None, 7, 10, 192) 576 conv2d_39[0][0]
__________________________________________________________________________________________________
activation_30 (Activation) (None, 7, 10, 192) 0 batch_normalization_30[0][0]
__________________________________________________________________________________________________
activation_33 (Activation) (None, 7, 10, 192) 0 batch_normalization_33[0][0]
__________________________________________________________________________________________________
activation_38 (Activation) (None, 7, 10, 192) 0 batch_normalization_38[0][0]
__________________________________________________________________________________________________
activation_39 (Activation) (None, 7, 10, 192) 0 batch_normalization_39[0][0]
__________________________________________________________________________________________________
mixed4 (Concatenate) (None, 7, 10, 768) 0 activation_30[0][0]
activation_33[0][0]
activation_38[0][0]
activation_39[0][0]
__________________________________________________________________________________________________
conv2d_44 (Conv2D) (None, 7, 10, 160) 122880 mixed4[0][0]
__________________________________________________________________________________________________
batch_normalization_44 (BatchNo (None, 7, 10, 160) 480 conv2d_44[0][0]
__________________________________________________________________________________________________
activation_44 (Activation) (None, 7, 10, 160) 0 batch_normalization_44[0][0]
__________________________________________________________________________________________________
conv2d_45 (Conv2D) (None, 7, 10, 160) 179200 activation_44[0][0]
__________________________________________________________________________________________________
batch_normalization_45 (BatchNo (None, 7, 10, 160) 480 conv2d_45[0][0]
__________________________________________________________________________________________________
activation_45 (Activation) (None, 7, 10, 160) 0 batch_normalization_45[0][0]
__________________________________________________________________________________________________
conv2d_41 (Conv2D) (None, 7, 10, 160) 122880 mixed4[0][0]
__________________________________________________________________________________________________
conv2d_46 (Conv2D) (None, 7, 10, 160) 179200 activation_45[0][0]
__________________________________________________________________________________________________
batch_normalization_41 (BatchNo (None, 7, 10, 160) 480 conv2d_41[0][0]
__________________________________________________________________________________________________
batch_normalization_46 (BatchNo (None, 7, 10, 160) 480 conv2d_46[0][0]
__________________________________________________________________________________________________
activation_41 (Activation) (None, 7, 10, 160) 0 batch_normalization_41[0][0]
__________________________________________________________________________________________________
activation_46 (Activation) (None, 7, 10, 160) 0 batch_normalization_46[0][0]
__________________________________________________________________________________________________
conv2d_42 (Conv2D) (None, 7, 10, 160) 179200 activation_41[0][0]
__________________________________________________________________________________________________
conv2d_47 (Conv2D) (None, 7, 10, 160) 179200 activation_46[0][0]
__________________________________________________________________________________________________
batch_normalization_42 (BatchNo (None, 7, 10, 160) 480 conv2d_42[0][0]
__________________________________________________________________________________________________
batch_normalization_47 (BatchNo (None, 7, 10, 160) 480 conv2d_47[0][0]
__________________________________________________________________________________________________
activation_42 (Activation) (None, 7, 10, 160) 0 batch_normalization_42[0][0]
__________________________________________________________________________________________________
activation_47 (Activation) (None, 7, 10, 160) 0 batch_normalization_47[0][0]
__________________________________________________________________________________________________
average_pooling2d_4 (AveragePoo (None, 7, 10, 768) 0 mixed4[0][0]
__________________________________________________________________________________________________
conv2d_40 (Conv2D) (None, 7, 10, 192) 147456 mixed4[0][0]
__________________________________________________________________________________________________
conv2d_43 (Conv2D) (None, 7, 10, 192) 215040 activation_42[0][0]
__________________________________________________________________________________________________
conv2d_48 (Conv2D) (None, 7, 10, 192) 215040 activation_47[0][0]
__________________________________________________________________________________________________
conv2d_49 (Conv2D) (None, 7, 10, 192) 147456 average_pooling2d_4[0][0]
__________________________________________________________________________________________________
batch_normalization_40 (BatchNo (None, 7, 10, 192) 576 conv2d_40[0][0]
__________________________________________________________________________________________________
batch_normalization_43 (BatchNo (None, 7, 10, 192) 576 conv2d_43[0][0]
__________________________________________________________________________________________________
batch_normalization_48 (BatchNo (None, 7, 10, 192) 576 conv2d_48[0][0]
__________________________________________________________________________________________________
batch_normalization_49 (BatchNo (None, 7, 10, 192) 576 conv2d_49[0][0]
__________________________________________________________________________________________________
activation_40 (Activation) (None, 7, 10, 192) 0 batch_normalization_40[0][0]
__________________________________________________________________________________________________
activation_43 (Activation) (None, 7, 10, 192) 0 batch_normalization_43[0][0]
__________________________________________________________________________________________________
activation_48 (Activation) (None, 7, 10, 192) 0 batch_normalization_48[0][0]
__________________________________________________________________________________________________
activation_49 (Activation) (None, 7, 10, 192) 0 batch_normalization_49[0][0]
__________________________________________________________________________________________________
mixed5 (Concatenate) (None, 7, 10, 768) 0 activation_40[0][0]
activation_43[0][0]
activation_48[0][0]
activation_49[0][0]
__________________________________________________________________________________________________
conv2d_54 (Conv2D) (None, 7, 10, 160) 122880 mixed5[0][0]
__________________________________________________________________________________________________
batch_normalization_54 (BatchNo (None, 7, 10, 160) 480 conv2d_54[0][0]
__________________________________________________________________________________________________
activation_54 (Activation) (None, 7, 10, 160) 0 batch_normalization_54[0][0]
__________________________________________________________________________________________________
conv2d_55 (Conv2D) (None, 7, 10, 160) 179200 activation_54[0][0]
__________________________________________________________________________________________________
batch_normalization_55 (BatchNo (None, 7, 10, 160) 480 conv2d_55[0][0]
__________________________________________________________________________________________________
activation_55 (Activation) (None, 7, 10, 160) 0 batch_normalization_55[0][0]
__________________________________________________________________________________________________
conv2d_51 (Conv2D) (None, 7, 10, 160) 122880 mixed5[0][0]
__________________________________________________________________________________________________
conv2d_56 (Conv2D) (None, 7, 10, 160) 179200 activation_55[0][0]
__________________________________________________________________________________________________
batch_normalization_51 (BatchNo (None, 7, 10, 160) 480 conv2d_51[0][0]
__________________________________________________________________________________________________
batch_normalization_56 (BatchNo (None, 7, 10, 160) 480 conv2d_56[0][0]
__________________________________________________________________________________________________
activation_51 (Activation) (None, 7, 10, 160) 0 batch_normalization_51[0][0]
__________________________________________________________________________________________________
activation_56 (Activation) (None, 7, 10, 160) 0 batch_normalization_56[0][0]
__________________________________________________________________________________________________
conv2d_52 (Conv2D) (None, 7, 10, 160) 179200 activation_51[0][0]
__________________________________________________________________________________________________
conv2d_57 (Conv2D) (None, 7, 10, 160) 179200 activation_56[0][0]
__________________________________________________________________________________________________
batch_normalization_52 (BatchNo (None, 7, 10, 160) 480 conv2d_52[0][0]
__________________________________________________________________________________________________
batch_normalization_57 (BatchNo (None, 7, 10, 160) 480 conv2d_57[0][0]
__________________________________________________________________________________________________
activation_52 (Activation) (None, 7, 10, 160) 0 batch_normalization_52[0][0]
__________________________________________________________________________________________________
activation_57 (Activation) (None, 7, 10, 160) 0 batch_normalization_57[0][0]
__________________________________________________________________________________________________
average_pooling2d_5 (AveragePoo (None, 7, 10, 768) 0 mixed5[0][0]
__________________________________________________________________________________________________
conv2d_50 (Conv2D) (None, 7, 10, 192) 147456 mixed5[0][0]
__________________________________________________________________________________________________
conv2d_53 (Conv2D) (None, 7, 10, 192) 215040 activation_52[0][0]
__________________________________________________________________________________________________
conv2d_58 (Conv2D) (None, 7, 10, 192) 215040 activation_57[0][0]
__________________________________________________________________________________________________
conv2d_59 (Conv2D) (None, 7, 10, 192) 147456 average_pooling2d_5[0][0]
__________________________________________________________________________________________________
batch_normalization_50 (BatchNo (None, 7, 10, 192) 576 conv2d_50[0][0]
__________________________________________________________________________________________________
batch_normalization_53 (BatchNo (None, 7, 10, 192) 576 conv2d_53[0][0]
__________________________________________________________________________________________________
batch_normalization_58 (BatchNo (None, 7, 10, 192) 576 conv2d_58[0][0]
__________________________________________________________________________________________________
batch_normalization_59 (BatchNo (None, 7, 10, 192) 576 conv2d_59[0][0]
__________________________________________________________________________________________________
activation_50 (Activation) (None, 7, 10, 192) 0 batch_normalization_50[0][0]
__________________________________________________________________________________________________
activation_53 (Activation) (None, 7, 10, 192) 0 batch_normalization_53[0][0]
__________________________________________________________________________________________________
activation_58 (Activation) (None, 7, 10, 192) 0 batch_normalization_58[0][0]
__________________________________________________________________________________________________
activation_59 (Activation) (None, 7, 10, 192) 0 batch_normalization_59[0][0]
__________________________________________________________________________________________________
mixed6 (Concatenate) (None, 7, 10, 768) 0 activation_50[0][0]
activation_53[0][0]
activation_58[0][0]
activation_59[0][0]
__________________________________________________________________________________________________
conv2d_64 (Conv2D) (None, 7, 10, 192) 147456 mixed6[0][0]
__________________________________________________________________________________________________
batch_normalization_64 (BatchNo (None, 7, 10, 192) 576 conv2d_64[0][0]
__________________________________________________________________________________________________
activation_64 (Activation) (None, 7, 10, 192) 0 batch_normalization_64[0][0]
__________________________________________________________________________________________________
conv2d_65 (Conv2D) (None, 7, 10, 192) 258048 activation_64[0][0]
__________________________________________________________________________________________________
batch_normalization_65 (BatchNo (None, 7, 10, 192) 576 conv2d_65[0][0]
__________________________________________________________________________________________________
activation_65 (Activation) (None, 7, 10, 192) 0 batch_normalization_65[0][0]
__________________________________________________________________________________________________
conv2d_61 (Conv2D) (None, 7, 10, 192) 147456 mixed6[0][0]
__________________________________________________________________________________________________
conv2d_66 (Conv2D) (None, 7, 10, 192) 258048 activation_65[0][0]
__________________________________________________________________________________________________
batch_normalization_61 (BatchNo (None, 7, 10, 192) 576 conv2d_61[0][0]
__________________________________________________________________________________________________
batch_normalization_66 (BatchNo (None, 7, 10, 192) 576 conv2d_66[0][0]
__________________________________________________________________________________________________
activation_61 (Activation) (None, 7, 10, 192) 0 batch_normalization_61[0][0]
__________________________________________________________________________________________________
activation_66 (Activation) (None, 7, 10, 192) 0 batch_normalization_66[0][0]
__________________________________________________________________________________________________
conv2d_62 (Conv2D) (None, 7, 10, 192) 258048 activation_61[0][0]
__________________________________________________________________________________________________
conv2d_67 (Conv2D) (None, 7, 10, 192) 258048 activation_66[0][0]
__________________________________________________________________________________________________
batch_normalization_62 (BatchNo (None, 7, 10, 192) 576 conv2d_62[0][0]
__________________________________________________________________________________________________
batch_normalization_67 (BatchNo (None, 7, 10, 192) 576 conv2d_67[0][0]
__________________________________________________________________________________________________
activation_62 (Activation) (None, 7, 10, 192) 0 batch_normalization_62[0][0]
__________________________________________________________________________________________________
activation_67 (Activation) (None, 7, 10, 192) 0 batch_normalization_67[0][0]
__________________________________________________________________________________________________
average_pooling2d_6 (AveragePoo (None, 7, 10, 768) 0 mixed6[0][0]
__________________________________________________________________________________________________
conv2d_60 (Conv2D) (None, 7, 10, 192) 147456 mixed6[0][0]
__________________________________________________________________________________________________
conv2d_63 (Conv2D) (None, 7, 10, 192) 258048 activation_62[0][0]
__________________________________________________________________________________________________
conv2d_68 (Conv2D) (None, 7, 10, 192) 258048 activation_67[0][0]
__________________________________________________________________________________________________
conv2d_69 (Conv2D) (None, 7, 10, 192) 147456 average_pooling2d_6[0][0]
__________________________________________________________________________________________________
batch_normalization_60 (BatchNo (None, 7, 10, 192) 576 conv2d_60[0][0]
__________________________________________________________________________________________________
batch_normalization_63 (BatchNo (None, 7, 10, 192) 576 conv2d_63[0][0]
__________________________________________________________________________________________________
batch_normalization_68 (BatchNo (None, 7, 10, 192) 576 conv2d_68[0][0]
__________________________________________________________________________________________________
batch_normalization_69 (BatchNo (None, 7, 10, 192) 576 conv2d_69[0][0]
__________________________________________________________________________________________________
activation_60 (Activation) (None, 7, 10, 192) 0 batch_normalization_60[0][0]
__________________________________________________________________________________________________
activation_63 (Activation) (None, 7, 10, 192) 0 batch_normalization_63[0][0]
__________________________________________________________________________________________________
activation_68 (Activation) (None, 7, 10, 192) 0 batch_normalization_68[0][0]
__________________________________________________________________________________________________
activation_69 (Activation) (None, 7, 10, 192) 0 batch_normalization_69[0][0]
__________________________________________________________________________________________________
mixed7 (Concatenate) (None, 7, 10, 768) 0 activation_60[0][0]
activation_63[0][0]
activation_68[0][0]
activation_69[0][0]
__________________________________________________________________________________________________
conv2d_72 (Conv2D) (None, 7, 10, 192) 147456 mixed7[0][0]
__________________________________________________________________________________________________
batch_normalization_72 (BatchNo (None, 7, 10, 192) 576 conv2d_72[0][0]
__________________________________________________________________________________________________
activation_72 (Activation) (None, 7, 10, 192) 0 batch_normalization_72[0][0]
__________________________________________________________________________________________________
conv2d_73 (Conv2D) (None, 7, 10, 192) 258048 activation_72[0][0]
__________________________________________________________________________________________________
batch_normalization_73 (BatchNo (None, 7, 10, 192) 576 conv2d_73[0][0]
__________________________________________________________________________________________________
activation_73 (Activation) (None, 7, 10, 192) 0 batch_normalization_73[0][0]
__________________________________________________________________________________________________
conv2d_70 (Conv2D) (None, 7, 10, 192) 147456 mixed7[0][0]
__________________________________________________________________________________________________
conv2d_74 (Conv2D) (None, 7, 10, 192) 258048 activation_73[0][0]
__________________________________________________________________________________________________
batch_normalization_70 (BatchNo (None, 7, 10, 192) 576 conv2d_70[0][0]
__________________________________________________________________________________________________
batch_normalization_74 (BatchNo (None, 7, 10, 192) 576 conv2d_74[0][0]
__________________________________________________________________________________________________
activation_70 (Activation) (None, 7, 10, 192) 0 batch_normalization_70[0][0]
__________________________________________________________________________________________________
activation_74 (Activation) (None, 7, 10, 192) 0 batch_normalization_74[0][0]
__________________________________________________________________________________________________
conv2d_71 (Conv2D) (None, 3, 4, 320) 552960 activation_70[0][0]
__________________________________________________________________________________________________
conv2d_75 (Conv2D) (None, 3, 4, 192) 331776 activation_74[0][0]
__________________________________________________________________________________________________
batch_normalization_71 (BatchNo (None, 3, 4, 320) 960 conv2d_71[0][0]
__________________________________________________________________________________________________
batch_normalization_75 (BatchNo (None, 3, 4, 192) 576 conv2d_75[0][0]
__________________________________________________________________________________________________
activation_71 (Activation) (None, 3, 4, 320) 0 batch_normalization_71[0][0]
__________________________________________________________________________________________________
activation_75 (Activation) (None, 3, 4, 192) 0 batch_normalization_75[0][0]
__________________________________________________________________________________________________
max_pooling2d_3 (MaxPooling2D) (None, 3, 4, 768) 0 mixed7[0][0]
__________________________________________________________________________________________________
mixed8 (Concatenate) (None, 3, 4, 1280) 0 activation_71[0][0]
activation_75[0][0]
max_pooling2d_3[0][0]
__________________________________________________________________________________________________
conv2d_80 (Conv2D) (None, 3, 4, 448) 573440 mixed8[0][0]
__________________________________________________________________________________________________
batch_normalization_80 (BatchNo (None, 3, 4, 448) 1344 conv2d_80[0][0]
__________________________________________________________________________________________________
activation_80 (Activation) (None, 3, 4, 448) 0 batch_normalization_80[0][0]
__________________________________________________________________________________________________
conv2d_77 (Conv2D) (None, 3, 4, 384) 491520 mixed8[0][0]
__________________________________________________________________________________________________
conv2d_81 (Conv2D) (None, 3, 4, 384) 1548288 activation_80[0][0]
__________________________________________________________________________________________________
batch_normalization_77 (BatchNo (None, 3, 4, 384) 1152 conv2d_77[0][0]
__________________________________________________________________________________________________
batch_normalization_81 (BatchNo (None, 3, 4, 384) 1152 conv2d_81[0][0]
__________________________________________________________________________________________________
activation_77 (Activation) (None, 3, 4, 384) 0 batch_normalization_77[0][0]
__________________________________________________________________________________________________
activation_81 (Activation) (None, 3, 4, 384) 0 batch_normalization_81[0][0]
__________________________________________________________________________________________________
conv2d_78 (Conv2D) (None, 3, 4, 384) 442368 activation_77[0][0]
__________________________________________________________________________________________________
conv2d_79 (Conv2D) (None, 3, 4, 384) 442368 activation_77[0][0]
__________________________________________________________________________________________________
conv2d_82 (Conv2D) (None, 3, 4, 384) 442368 activation_81[0][0]
__________________________________________________________________________________________________
conv2d_83 (Conv2D) (None, 3, 4, 384) 442368 activation_81[0][0]
__________________________________________________________________________________________________
average_pooling2d_7 (AveragePoo (None, 3, 4, 1280) 0 mixed8[0][0]
__________________________________________________________________________________________________
conv2d_76 (Conv2D) (None, 3, 4, 320) 409600 mixed8[0][0]
__________________________________________________________________________________________________
batch_normalization_78 (BatchNo (None, 3, 4, 384) 1152 conv2d_78[0][0]
__________________________________________________________________________________________________
batch_normalization_79 (BatchNo (None, 3, 4, 384) 1152 conv2d_79[0][0]
__________________________________________________________________________________________________
batch_normalization_82 (BatchNo (None, 3, 4, 384) 1152 conv2d_82[0][0]
__________________________________________________________________________________________________
batch_normalization_83 (BatchNo (None, 3, 4, 384) 1152 conv2d_83[0][0]
__________________________________________________________________________________________________
conv2d_84 (Conv2D) (None, 3, 4, 192) 245760 average_pooling2d_7[0][0]
__________________________________________________________________________________________________
batch_normalization_76 (BatchNo (None, 3, 4, 320) 960 conv2d_76[0][0]
__________________________________________________________________________________________________
activation_78 (Activation) (None, 3, 4, 384) 0 batch_normalization_78[0][0]
__________________________________________________________________________________________________
activation_79 (Activation) (None, 3, 4, 384) 0 batch_normalization_79[0][0]
__________________________________________________________________________________________________
activation_82 (Activation) (None, 3, 4, 384) 0 batch_normalization_82[0][0]
__________________________________________________________________________________________________
activation_83 (Activation) (None, 3, 4, 384) 0 batch_normalization_83[0][0]
__________________________________________________________________________________________________
batch_normalization_84 (BatchNo (None, 3, 4, 192) 576 conv2d_84[0][0]
__________________________________________________________________________________________________
activation_76 (Activation) (None, 3, 4, 320) 0 batch_normalization_76[0][0]
__________________________________________________________________________________________________
mixed9_0 (Concatenate) (None, 3, 4, 768) 0 activation_78[0][0]
activation_79[0][0]
__________________________________________________________________________________________________
concatenate (Concatenate) (None, 3, 4, 768) 0 activation_82[0][0]
activation_83[0][0]
__________________________________________________________________________________________________
activation_84 (Activation) (None, 3, 4, 192) 0 batch_normalization_84[0][0]
__________________________________________________________________________________________________
mixed9 (Concatenate) (None, 3, 4, 2048) 0 activation_76[0][0]
mixed9_0[0][0]
concatenate[0][0]
activation_84[0][0]
__________________________________________________________________________________________________
conv2d_89 (Conv2D) (None, 3, 4, 448) 917504 mixed9[0][0]
__________________________________________________________________________________________________
batch_normalization_89 (BatchNo (None, 3, 4, 448) 1344 conv2d_89[0][0]
__________________________________________________________________________________________________
activation_89 (Activation) (None, 3, 4, 448) 0 batch_normalization_89[0][0]
__________________________________________________________________________________________________
conv2d_86 (Conv2D) (None, 3, 4, 384) 786432 mixed9[0][0]
__________________________________________________________________________________________________
conv2d_90 (Conv2D) (None, 3, 4, 384) 1548288 activation_89[0][0]
__________________________________________________________________________________________________
batch_normalization_86 (BatchNo (None, 3, 4, 384) 1152 conv2d_86[0][0]
__________________________________________________________________________________________________
batch_normalization_90 (BatchNo (None, 3, 4, 384) 1152 conv2d_90[0][0]
__________________________________________________________________________________________________
activation_86 (Activation) (None, 3, 4, 384) 0 batch_normalization_86[0][0]
__________________________________________________________________________________________________
activation_90 (Activation) (None, 3, 4, 384) 0 batch_normalization_90[0][0]
__________________________________________________________________________________________________
conv2d_87 (Conv2D) (None, 3, 4, 384) 442368 activation_86[0][0]
__________________________________________________________________________________________________
conv2d_88 (Conv2D) (None, 3, 4, 384) 442368 activation_86[0][0]
__________________________________________________________________________________________________
conv2d_91 (Conv2D) (None, 3, 4, 384) 442368 activation_90[0][0]
__________________________________________________________________________________________________
conv2d_92 (Conv2D) (None, 3, 4, 384) 442368 activation_90[0][0]
__________________________________________________________________________________________________
average_pooling2d_8 (AveragePoo (None, 3, 4, 2048) 0 mixed9[0][0]
__________________________________________________________________________________________________
conv2d_85 (Conv2D) (None, 3, 4, 320) 655360 mixed9[0][0]
__________________________________________________________________________________________________
batch_normalization_87 (BatchNo (None, 3, 4, 384) 1152 conv2d_87[0][0]
__________________________________________________________________________________________________
batch_normalization_88 (BatchNo (None, 3, 4, 384) 1152 conv2d_88[0][0]
__________________________________________________________________________________________________
batch_normalization_91 (BatchNo (None, 3, 4, 384) 1152 conv2d_91[0][0]
__________________________________________________________________________________________________
batch_normalization_92 (BatchNo (None, 3, 4, 384) 1152 conv2d_92[0][0]
__________________________________________________________________________________________________
conv2d_93 (Conv2D) (None, 3, 4, 192) 393216 average_pooling2d_8[0][0]
__________________________________________________________________________________________________
batch_normalization_85 (BatchNo (None, 3, 4, 320) 960 conv2d_85[0][0]
__________________________________________________________________________________________________
activation_87 (Activation) (None, 3, 4, 384) 0 batch_normalization_87[0][0]
__________________________________________________________________________________________________
activation_88 (Activation) (None, 3, 4, 384) 0 batch_normalization_88[0][0]
__________________________________________________________________________________________________
activation_91 (Activation) (None, 3, 4, 384) 0 batch_normalization_91[0][0]
__________________________________________________________________________________________________
activation_92 (Activation) (None, 3, 4, 384) 0 batch_normalization_92[0][0]
__________________________________________________________________________________________________
batch_normalization_93 (BatchNo (None, 3, 4, 192) 576 conv2d_93[0][0]
__________________________________________________________________________________________________
activation_85 (Activation) (None, 3, 4, 320) 0 batch_normalization_85[0][0]
__________________________________________________________________________________________________
mixed9_1 (Concatenate) (None, 3, 4, 768) 0 activation_87[0][0]
activation_88[0][0]
__________________________________________________________________________________________________
concatenate_1 (Concatenate) (None, 3, 4, 768) 0 activation_91[0][0]
activation_92[0][0]
__________________________________________________________________________________________________
activation_93 (Activation) (None, 3, 4, 192) 0 batch_normalization_93[0][0]
__________________________________________________________________________________________________
mixed10 (Concatenate) (None, 3, 4, 2048) 0 activation_85[0][0]
mixed9_1[0][0]
concatenate_1[0][0]
activation_93[0][0]
__________________________________________________________________________________________________
flatten (Flatten) (None, 24576) 0 mixed10[0][0]
__________________________________________________________________________________________________
dense (Dense) (None, 1024) 25166848 flatten[0][0]
__________________________________________________________________________________________________
dense_6 (Dense) (None, 1024) 25166848 flatten[0][0]
__________________________________________________________________________________________________
dense_12 (Dense) (None, 1024) 25166848 flatten[0][0]
__________________________________________________________________________________________________
activation_94 (Activation) (None, 1024) 0 dense[0][0]
__________________________________________________________________________________________________
activation_99 (Activation) (None, 1024) 0 dense_6[0][0]
__________________________________________________________________________________________________
activation_104 (Activation) (None, 1024) 0 dense_12[0][0]
__________________________________________________________________________________________________
batch_normalization_94 (BatchNo (None, 1024) 4096 activation_94[0][0]
__________________________________________________________________________________________________
batch_normalization_99 (BatchNo (None, 1024) 4096 activation_99[0][0]
__________________________________________________________________________________________________
batch_normalization_104 (BatchN (None, 1024) 4096 activation_104[0][0]
__________________________________________________________________________________________________
dropout (Dropout) (None, 1024) 0 batch_normalization_94[0][0]
__________________________________________________________________________________________________
dropout_5 (Dropout) (None, 1024) 0 batch_normalization_99[0][0]
__________________________________________________________________________________________________
dropout_10 (Dropout) (None, 1024) 0 batch_normalization_104[0][0]
__________________________________________________________________________________________________
dense_1 (Dense) (None, 512) 524800 dropout[0][0]
__________________________________________________________________________________________________
dense_7 (Dense) (None, 512) 524800 dropout_5[0][0]
__________________________________________________________________________________________________
dense_13 (Dense) (None, 512) 524800 dropout_10[0][0]
__________________________________________________________________________________________________
activation_95 (Activation) (None, 512) 0 dense_1[0][0]
__________________________________________________________________________________________________
activation_100 (Activation) (None, 512) 0 dense_7[0][0]
__________________________________________________________________________________________________
activation_105 (Activation) (None, 512) 0 dense_13[0][0]
__________________________________________________________________________________________________
batch_normalization_95 (BatchNo (None, 512) 2048 activation_95[0][0]
__________________________________________________________________________________________________
batch_normalization_100 (BatchN (None, 512) 2048 activation_100[0][0]
__________________________________________________________________________________________________
batch_normalization_105 (BatchN (None, 512) 2048 activation_105[0][0]
__________________________________________________________________________________________________
dropout_1 (Dropout) (None, 512) 0 batch_normalization_95[0][0]
__________________________________________________________________________________________________
dropout_6 (Dropout) (None, 512) 0 batch_normalization_100[0][0]
__________________________________________________________________________________________________
dropout_11 (Dropout) (None, 512) 0 batch_normalization_105[0][0]
__________________________________________________________________________________________________
dense_2 (Dense) (None, 256) 131328 dropout_1[0][0]
__________________________________________________________________________________________________
dense_8 (Dense) (None, 256) 131328 dropout_6[0][0]
__________________________________________________________________________________________________
dense_14 (Dense) (None, 256) 131328 dropout_11[0][0]
__________________________________________________________________________________________________
activation_96 (Activation) (None, 256) 0 dense_2[0][0]
__________________________________________________________________________________________________
activation_101 (Activation) (None, 256) 0 dense_8[0][0]
__________________________________________________________________________________________________
activation_106 (Activation) (None, 256) 0 dense_14[0][0]
__________________________________________________________________________________________________
batch_normalization_96 (BatchNo (None, 256) 1024 activation_96[0][0]
__________________________________________________________________________________________________
batch_normalization_101 (BatchN (None, 256) 1024 activation_101[0][0]
__________________________________________________________________________________________________
batch_normalization_106 (BatchN (None, 256) 1024 activation_106[0][0]
__________________________________________________________________________________________________
dropout_2 (Dropout) (None, 256) 0 batch_normalization_96[0][0]
__________________________________________________________________________________________________
dropout_7 (Dropout) (None, 256) 0 batch_normalization_101[0][0]
__________________________________________________________________________________________________
dropout_12 (Dropout) (None, 256) 0 batch_normalization_106[0][0]
__________________________________________________________________________________________________
dense_3 (Dense) (None, 128) 32896 dropout_2[0][0]
__________________________________________________________________________________________________
dense_9 (Dense) (None, 128) 32896 dropout_7[0][0]
__________________________________________________________________________________________________
dense_15 (Dense) (None, 128) 32896 dropout_12[0][0]
__________________________________________________________________________________________________
activation_97 (Activation) (None, 128) 0 dense_3[0][0]
__________________________________________________________________________________________________
activation_102 (Activation) (None, 128) 0 dense_9[0][0]
__________________________________________________________________________________________________
activation_107 (Activation) (None, 128) 0 dense_15[0][0]
__________________________________________________________________________________________________
batch_normalization_97 (BatchNo (None, 128) 512 activation_97[0][0]
__________________________________________________________________________________________________
batch_normalization_102 (BatchN (None, 128) 512 activation_102[0][0]
__________________________________________________________________________________________________
batch_normalization_107 (BatchN (None, 128) 512 activation_107[0][0]
__________________________________________________________________________________________________
dropout_3 (Dropout) (None, 128) 0 batch_normalization_97[0][0]
__________________________________________________________________________________________________
dropout_8 (Dropout) (None, 128) 0 batch_normalization_102[0][0]
__________________________________________________________________________________________________
dropout_13 (Dropout) (None, 128) 0 batch_normalization_107[0][0]
__________________________________________________________________________________________________
dense_4 (Dense) (None, 128) 16512 dropout_3[0][0]
__________________________________________________________________________________________________
dense_10 (Dense) (None, 128) 16512 dropout_8[0][0]
__________________________________________________________________________________________________
dense_16 (Dense) (None, 128) 16512 dropout_13[0][0]
__________________________________________________________________________________________________
activation_98 (Activation) (None, 128) 0 dense_4[0][0]
__________________________________________________________________________________________________
activation_103 (Activation) (None, 128) 0 dense_10[0][0]
__________________________________________________________________________________________________
activation_108 (Activation) (None, 128) 0 dense_16[0][0]
__________________________________________________________________________________________________
batch_normalization_98 (BatchNo (None, 128) 512 activation_98[0][0]
__________________________________________________________________________________________________
batch_normalization_103 (BatchN (None, 128) 512 activation_103[0][0]
__________________________________________________________________________________________________
batch_normalization_108 (BatchN (None, 128) 512 activation_108[0][0]
__________________________________________________________________________________________________
dropout_4 (Dropout) (None, 128) 0 batch_normalization_98[0][0]
__________________________________________________________________________________________________
dropout_9 (Dropout) (None, 128) 0 batch_normalization_103[0][0]
__________________________________________________________________________________________________
dropout_14 (Dropout) (None, 128) 0 batch_normalization_108[0][0]
__________________________________________________________________________________________________
dense_5 (Dense) (None, 6) 774 dropout_4[0][0]
__________________________________________________________________________________________________
dense_11 (Dense) (None, 1) 129 dropout_9[0][0]
__________________________________________________________________________________________________
dense_17 (Dense) (None, 1) 129 dropout_14[0][0]
__________________________________________________________________________________________________
Genre (Activation) (None, 6) 0 dense_5[0][0]
__________________________________________________________________________________________________
Rating (Activation) (None, 1) 0 dense_11[0][0]
__________________________________________________________________________________________________
Year (Activation) (None, 1) 0 dense_17[0][0]
==================================================================================================
Total params: 99,445,544
Trainable params: 77,630,472
Non-trainable params: 21,815,072
__________________________________________________________________________________________________
from tensorflow.keras.callbacks import ModelCheckpoint
from tensorflow.keras.backend import clear_session
import tensorflow as tf
config = tf.compat.v1.ConfigProto()
config.gpu_options.allow_growth = True
sess = tf.compat.v1.Session(config=config)
tf.compat.v1.keras.backend.set_session(sess)
# GPU check - CUDA 11.1 with cuDNN 8.05 on 460.39 - GTX 1060 6GB
print(tf.test.is_built_with_cuda())
print(tf.config.list_physical_devices('GPU'))
clear_session()
True [PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]
tr_generator = data_generator.generate_images(train_idx, cfgs.HYPERPARAMS.BATCH_SIZE_TR, True)
val_generator = data_generator.generate_images(valid_idx, cfgs.HYPERPARAMS.BATCH_SIZE_VAL, True)
import datetime
%load_ext tensorboard
log_dir = log_dir = "./logs/" + datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
callbacks = [keras.callbacks.TensorBoard(log_dir=log_dir, histogram_freq=0)]
# keras.callbacks.ModelCheckpoint('./checkpoints', monitor='val_loss', verbose=0,
# save_best_only=False, save_weights_only=False,
# mode='auto', save_freq='epoch')
%%time
history = content_model.model.fit(tr_generator,
steps_per_epoch=len(train_idx)//cfgs.HYPERPARAMS.BATCH_SIZE_TR,
epochs=cfgs.HYPERPARAMS.EPOCHS,
callbacks=callbacks,
validation_data=val_generator,
validation_steps=len(valid_idx)//cfgs.HYPERPARAMS.BATCH_SIZE_VAL)
Epoch 1/30 126/151 [========================>.....] - ETA: 3s - loss: 11.9191 - Genre_loss: 2.8610 - Rating_loss: 4.1616 - Year_loss: 4.8966 - Genre_accuracy: 0.1792 - Rating_mae: 1.6071 - Year_mae: 1.7484 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 29s 148ms/step - loss: 11.7802 - Genre_loss: 2.8458 - Rating_loss: 4.0894 - Year_loss: 4.8451 - Genre_accuracy: 0.1804 - Rating_mae: 1.5912 - Year_mae: 1.7380 - val_loss: 3.5469 - val_Genre_loss: 1.7516 - val_Rating_loss: 0.8759 - val_Year_loss: 0.9194 - val_Genre_accuracy: 0.2861 - val_Rating_mae: 0.8235 - val_Year_mae: 0.9069 Epoch 2/30 127/151 [========================>.....] - ETA: 3s - loss: 8.8550 - Genre_loss: 2.6069 - Rating_loss: 2.8316 - Year_loss: 3.4164 - Genre_accuracy: 0.2071 - Rating_mae: 1.3197 - Year_mae: 1.4476 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 140ms/step - loss: 8.8025 - Genre_loss: 2.6000 - Rating_loss: 2.8045 - Year_loss: 3.3980 - Genre_accuracy: 0.2082 - Rating_mae: 1.3134 - Year_mae: 1.4444 - val_loss: 2.3532 - val_Genre_loss: 1.6882 - val_Rating_loss: 0.1931 - val_Year_loss: 0.4719 - val_Genre_accuracy: 0.3213 - val_Rating_mae: 0.3611 - val_Year_mae: 0.6319 Epoch 3/30 128/151 [========================>.....] - ETA: 2s - loss: 7.6189 - Genre_loss: 2.4831 - Rating_loss: 2.3767 - Year_loss: 2.7591 - Genre_accuracy: 0.2238 - Rating_mae: 1.2016 - Year_mae: 1.3047 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 141ms/step - loss: 7.5629 - Genre_loss: 2.4763 - Rating_loss: 2.3395 - Year_loss: 2.7470 - Genre_accuracy: 0.2251 - Rating_mae: 1.1928 - Year_mae: 1.3016 - val_loss: 2.1156 - val_Genre_loss: 1.6385 - val_Rating_loss: 0.1716 - val_Year_loss: 0.3056 - val_Genre_accuracy: 0.3496 - val_Rating_mae: 0.3507 - val_Year_mae: 0.4877 Epoch 4/30 129/151 [========================>.....] - ETA: 2s - loss: 6.5935 - Genre_loss: 2.3251 - Rating_loss: 1.8703 - Year_loss: 2.3982 - Genre_accuracy: 0.2449 - Rating_mae: 1.0745 - Year_mae: 1.2205 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 143ms/step - loss: 6.5523 - Genre_loss: 2.3212 - Rating_loss: 1.8566 - Year_loss: 2.3744 - Genre_accuracy: 0.2445 - Rating_mae: 1.0701 - Year_mae: 1.2149 - val_loss: 1.8989 - val_Genre_loss: 1.5745 - val_Rating_loss: 0.1435 - val_Year_loss: 0.1809 - val_Genre_accuracy: 0.3633 - val_Rating_mae: 0.3174 - val_Year_mae: 0.3689 Epoch 5/30 130/151 [========================>.....] - ETA: 2s - loss: 5.7569 - Genre_loss: 2.2363 - Rating_loss: 1.5969 - Year_loss: 1.9237 - Genre_accuracy: 0.2402 - Rating_mae: 0.9861 - Year_mae: 1.0936 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 145ms/step - loss: 5.7338 - Genre_loss: 2.2314 - Rating_loss: 1.5843 - Year_loss: 1.9181 - Genre_accuracy: 0.2423 - Rating_mae: 0.9823 - Year_mae: 1.0913 - val_loss: 1.8418 - val_Genre_loss: 1.5383 - val_Rating_loss: 0.1304 - val_Year_loss: 0.1731 - val_Genre_accuracy: 0.3926 - val_Rating_mae: 0.2999 - val_Year_mae: 0.3683 Epoch 6/30 131/151 [=========================>....] - ETA: 2s - loss: 5.3260 - Genre_loss: 2.1599 - Rating_loss: 1.3875 - Year_loss: 1.7786 - Genre_accuracy: 0.2641 - Rating_mae: 0.9238 - Year_mae: 1.0420 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 146ms/step - loss: 5.3012 - Genre_loss: 2.1552 - Rating_loss: 1.3779 - Year_loss: 1.7681 - Genre_accuracy: 0.2659 - Rating_mae: 0.9205 - Year_mae: 1.0389 - val_loss: 1.7060 - val_Genre_loss: 1.5154 - val_Rating_loss: 0.0993 - val_Year_loss: 0.0913 - val_Genre_accuracy: 0.3945 - val_Rating_mae: 0.2612 - val_Year_mae: 0.2540 Epoch 7/30 132/151 [=========================>....] - ETA: 2s - loss: 4.8579 - Genre_loss: 2.0925 - Rating_loss: 1.1909 - Year_loss: 1.5745 - Genre_accuracy: 0.2769 - Rating_mae: 0.8583 - Year_mae: 0.9822 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 146ms/step - loss: 4.8390 - Genre_loss: 2.0885 - Rating_loss: 1.1864 - Year_loss: 1.5642 - Genre_accuracy: 0.2773 - Rating_mae: 0.8557 - Year_mae: 0.9788 - val_loss: 1.6567 - val_Genre_loss: 1.4916 - val_Rating_loss: 0.1045 - val_Year_loss: 0.0607 - val_Genre_accuracy: 0.4111 - val_Rating_mae: 0.2732 - val_Year_mae: 0.2024 Epoch 8/30 132/151 [=========================>....] - ETA: 2s - loss: 4.3968 - Genre_loss: 2.0042 - Rating_loss: 1.0310 - Year_loss: 1.3616 - Genre_accuracy: 0.2917 - Rating_mae: 0.7912 - Year_mae: 0.9195 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 146ms/step - loss: 4.3908 - Genre_loss: 2.0006 - Rating_loss: 1.0305 - Year_loss: 1.3596 - Genre_accuracy: 0.2928 - Rating_mae: 0.7911 - Year_mae: 0.9183 - val_loss: 1.6267 - val_Genre_loss: 1.4776 - val_Rating_loss: 0.0952 - val_Year_loss: 0.0539 - val_Genre_accuracy: 0.4307 - val_Rating_mae: 0.2600 - val_Year_mae: 0.1907 Epoch 9/30 133/151 [=========================>....] - ETA: 2s - loss: 4.0410 - Genre_loss: 1.9228 - Rating_loss: 0.9210 - Year_loss: 1.1972 - Genre_accuracy: 0.3188 - Rating_mae: 0.7520 - Year_mae: 0.8618 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 147ms/step - loss: 4.0322 - Genre_loss: 1.9208 - Rating_loss: 0.9168 - Year_loss: 1.1946 - Genre_accuracy: 0.3191 - Rating_mae: 0.7498 - Year_mae: 0.8607 - val_loss: 1.5822 - val_Genre_loss: 1.4627 - val_Rating_loss: 0.0677 - val_Year_loss: 0.0517 - val_Genre_accuracy: 0.4336 - val_Rating_mae: 0.2135 - val_Year_mae: 0.1903 Epoch 10/30 134/151 [=========================>....] - ETA: 2s - loss: 3.9318 - Genre_loss: 1.9099 - Rating_loss: 0.8568 - Year_loss: 1.1651 - Genre_accuracy: 0.2977 - Rating_mae: 0.7191 - Year_mae: 0.8340 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 148ms/step - loss: 3.9136 - Genre_loss: 1.9070 - Rating_loss: 0.8508 - Year_loss: 1.1558 - Genre_accuracy: 0.2988 - Rating_mae: 0.7171 - Year_mae: 0.8314 - val_loss: 1.5617 - val_Genre_loss: 1.4637 - val_Rating_loss: 0.0517 - val_Year_loss: 0.0463 - val_Genre_accuracy: 0.4346 - val_Rating_mae: 0.1865 - val_Year_mae: 0.1756 Epoch 11/30 135/151 [=========================>....] - ETA: 2s - loss: 3.5683 - Genre_loss: 1.8244 - Rating_loss: 0.7507 - Year_loss: 0.9931 - Genre_accuracy: 0.3253 - Rating_mae: 0.6711 - Year_mae: 0.7783 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 149ms/step - loss: 3.5592 - Genre_loss: 1.8234 - Rating_loss: 0.7469 - Year_loss: 0.9889 - Genre_accuracy: 0.3256 - Rating_mae: 0.6700 - Year_mae: 0.7767 - val_loss: 1.5308 - val_Genre_loss: 1.4374 - val_Rating_loss: 0.0632 - val_Year_loss: 0.0302 - val_Genre_accuracy: 0.4395 - val_Rating_mae: 0.2082 - val_Year_mae: 0.1426 Epoch 12/30 136/151 [==========================>...] - ETA: 2s - loss: 3.3105 - Genre_loss: 1.8020 - Rating_loss: 0.6241 - Year_loss: 0.8843 - Genre_accuracy: 0.3284 - Rating_mae: 0.6176 - Year_mae: 0.7369 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 23s 149ms/step - loss: 3.3065 - Genre_loss: 1.8005 - Rating_loss: 0.6241 - Year_loss: 0.8818 - Genre_accuracy: 0.3291 - Rating_mae: 0.6177 - Year_mae: 0.7357 - val_loss: 1.5199 - val_Genre_loss: 1.4357 - val_Rating_loss: 0.0540 - val_Year_loss: 0.0302 - val_Genre_accuracy: 0.4365 - val_Rating_mae: 0.1925 - val_Year_mae: 0.1419 Epoch 13/30 137/151 [==========================>...] - ETA: 1s - loss: 3.0877 - Genre_loss: 1.7607 - Rating_loss: 0.5579 - Year_loss: 0.7691 - Genre_accuracy: 0.3441 - Rating_mae: 0.5808 - Year_mae: 0.6844 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 149ms/step - loss: 3.0826 - Genre_loss: 1.7583 - Rating_loss: 0.5575 - Year_loss: 0.7668 - Genre_accuracy: 0.3442 - Rating_mae: 0.5808 - Year_mae: 0.6836 - val_loss: 1.4876 - val_Genre_loss: 1.4185 - val_Rating_loss: 0.0431 - val_Year_loss: 0.0260 - val_Genre_accuracy: 0.4561 - val_Rating_mae: 0.1716 - val_Year_mae: 0.1301 Epoch 14/30 138/151 [==========================>...] - ETA: 1s - loss: 2.9234 - Genre_loss: 1.7177 - Rating_loss: 0.5120 - Year_loss: 0.6936 - Genre_accuracy: 0.3586 - Rating_mae: 0.5612 - Year_mae: 0.6503 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 140ms/step - loss: 2.9183 - Genre_loss: 1.7149 - Rating_loss: 0.5109 - Year_loss: 0.6925 - Genre_accuracy: 0.3593 - Rating_mae: 0.5605 - Year_mae: 0.6498 - val_loss: 1.4888 - val_Genre_loss: 1.4114 - val_Rating_loss: 0.0482 - val_Year_loss: 0.0292 - val_Genre_accuracy: 0.4551 - val_Rating_mae: 0.1827 - val_Year_mae: 0.1421 Epoch 15/30 138/151 [==========================>...] - ETA: 1s - loss: 2.7295 - Genre_loss: 1.6662 - Rating_loss: 0.4410 - Year_loss: 0.6223 - Genre_accuracy: 0.3645 - Rating_mae: 0.5186 - Year_mae: 0.6197 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 139ms/step - loss: 2.7248 - Genre_loss: 1.6646 - Rating_loss: 0.4397 - Year_loss: 0.6205 - Genre_accuracy: 0.3652 - Rating_mae: 0.5178 - Year_mae: 0.6190 - val_loss: 1.5013 - val_Genre_loss: 1.4217 - val_Rating_loss: 0.0520 - val_Year_loss: 0.0276 - val_Genre_accuracy: 0.4590 - val_Rating_mae: 0.1923 - val_Year_mae: 0.1402 Epoch 16/30 139/151 [==========================>...] - ETA: 1s - loss: 2.6024 - Genre_loss: 1.6462 - Rating_loss: 0.3859 - Year_loss: 0.5703 - Genre_accuracy: 0.3755 - Rating_mae: 0.4823 - Year_mae: 0.5910 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 141ms/step - loss: 2.5976 - Genre_loss: 1.6444 - Rating_loss: 0.3849 - Year_loss: 0.5683 - Genre_accuracy: 0.3757 - Rating_mae: 0.4818 - Year_mae: 0.5902 - val_loss: 1.4714 - val_Genre_loss: 1.4030 - val_Rating_loss: 0.0403 - val_Year_loss: 0.0282 - val_Genre_accuracy: 0.4717 - val_Rating_mae: 0.1666 - val_Year_mae: 0.1433 Epoch 17/30 140/151 [==========================>...] - ETA: 1s - loss: 2.4575 - Genre_loss: 1.6080 - Rating_loss: 0.3463 - Year_loss: 0.5032 - Genre_accuracy: 0.3847 - Rating_mae: 0.4608 - Year_mae: 0.5572 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 141ms/step - loss: 2.4541 - Genre_loss: 1.6066 - Rating_loss: 0.3457 - Year_loss: 0.5019 - Genre_accuracy: 0.3851 - Rating_mae: 0.4603 - Year_mae: 0.5565 - val_loss: 1.4654 - val_Genre_loss: 1.3990 - val_Rating_loss: 0.0404 - val_Year_loss: 0.0260 - val_Genre_accuracy: 0.4717 - val_Rating_mae: 0.1668 - val_Year_mae: 0.1375 Epoch 18/30 141/151 [===========================>..] - ETA: 1s - loss: 2.3234 - Genre_loss: 1.5874 - Rating_loss: 0.2991 - Year_loss: 0.4369 - Genre_accuracy: 0.3861 - Rating_mae: 0.4302 - Year_mae: 0.5187 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 142ms/step - loss: 2.3209 - Genre_loss: 1.5862 - Rating_loss: 0.2986 - Year_loss: 0.4362 - Genre_accuracy: 0.3869 - Rating_mae: 0.4298 - Year_mae: 0.5182 - val_loss: 1.4463 - val_Genre_loss: 1.3815 - val_Rating_loss: 0.0392 - val_Year_loss: 0.0256 - val_Genre_accuracy: 0.4775 - val_Rating_mae: 0.1665 - val_Year_mae: 0.1375 Epoch 19/30 142/151 [===========================>..] - ETA: 1s - loss: 2.2176 - Genre_loss: 1.5540 - Rating_loss: 0.2764 - Year_loss: 0.3872 - Genre_accuracy: 0.4076 - Rating_mae: 0.4067 - Year_mae: 0.4885 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 139ms/step - loss: 2.2156 - Genre_loss: 1.5535 - Rating_loss: 0.2752 - Year_loss: 0.3869 - Genre_accuracy: 0.4076 - Rating_mae: 0.4060 - Year_mae: 0.4881 - val_loss: 1.4411 - val_Genre_loss: 1.3824 - val_Rating_loss: 0.0381 - val_Year_loss: 0.0205 - val_Genre_accuracy: 0.4805 - val_Rating_mae: 0.1648 - val_Year_mae: 0.1221 Epoch 20/30 143/151 [===========================>..] - ETA: 1s - loss: 2.1003 - Genre_loss: 1.5236 - Rating_loss: 0.2314 - Year_loss: 0.3453 - Genre_accuracy: 0.4158 - Rating_mae: 0.3768 - Year_mae: 0.4580 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 139ms/step - loss: 2.0996 - Genre_loss: 1.5234 - Rating_loss: 0.2311 - Year_loss: 0.3452 - Genre_accuracy: 0.4157 - Rating_mae: 0.3766 - Year_mae: 0.4580 - val_loss: 1.4343 - val_Genre_loss: 1.3739 - val_Rating_loss: 0.0353 - val_Year_loss: 0.0251 - val_Genre_accuracy: 0.4863 - val_Rating_mae: 0.1583 - val_Year_mae: 0.1392 Epoch 21/30 144/151 [===========================>..] - ETA: 0s - loss: 2.0047 - Genre_loss: 1.4931 - Rating_loss: 0.1988 - Year_loss: 0.3128 - Genre_accuracy: 0.4154 - Rating_mae: 0.3514 - Year_mae: 0.4382 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 140ms/step - loss: 2.0035 - Genre_loss: 1.4927 - Rating_loss: 0.1984 - Year_loss: 0.3124 - Genre_accuracy: 0.4159 - Rating_mae: 0.3510 - Year_mae: 0.4379 - val_loss: 1.4253 - val_Genre_loss: 1.3655 - val_Rating_loss: 0.0346 - val_Year_loss: 0.0252 - val_Genre_accuracy: 0.4746 - val_Rating_mae: 0.1569 - val_Year_mae: 0.1419 Epoch 22/30 144/151 [===========================>..] - ETA: 0s - loss: 1.9072 - Genre_loss: 1.4452 - Rating_loss: 0.1760 - Year_loss: 0.2860 - Genre_accuracy: 0.4326 - Rating_mae: 0.3305 - Year_mae: 0.4201 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 139ms/step - loss: 1.9064 - Genre_loss: 1.4450 - Rating_loss: 0.1759 - Year_loss: 0.2856 - Genre_accuracy: 0.4330 - Rating_mae: 0.3304 - Year_mae: 0.4197 - val_loss: 1.4145 - val_Genre_loss: 1.3558 - val_Rating_loss: 0.0351 - val_Year_loss: 0.0236 - val_Genre_accuracy: 0.4893 - val_Rating_mae: 0.1583 - val_Year_mae: 0.1372 Epoch 23/30 145/151 [===========================>..] - ETA: 0s - loss: 1.8555 - Genre_loss: 1.4474 - Rating_loss: 0.1577 - Year_loss: 0.2504 - Genre_accuracy: 0.4454 - Rating_mae: 0.3133 - Year_mae: 0.3938 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 140ms/step - loss: 1.8546 - Genre_loss: 1.4468 - Rating_loss: 0.1575 - Year_loss: 0.2502 - Genre_accuracy: 0.4454 - Rating_mae: 0.3130 - Year_mae: 0.3936 - val_loss: 1.4123 - val_Genre_loss: 1.3569 - val_Rating_loss: 0.0329 - val_Year_loss: 0.0225 - val_Genre_accuracy: 0.4932 - val_Rating_mae: 0.1534 - val_Year_mae: 0.1353 Epoch 24/30 146/151 [============================>.] - ETA: 0s - loss: 1.7781 - Genre_loss: 1.4061 - Rating_loss: 0.1439 - Year_loss: 0.2282 - Genre_accuracy: 0.4523 - Rating_mae: 0.2983 - Year_mae: 0.3748 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 140ms/step - loss: 1.7770 - Genre_loss: 1.4056 - Rating_loss: 0.1436 - Year_loss: 0.2277 - Genre_accuracy: 0.4526 - Rating_mae: 0.2980 - Year_mae: 0.3744 - val_loss: 1.4112 - val_Genre_loss: 1.3595 - val_Rating_loss: 0.0318 - val_Year_loss: 0.0200 - val_Genre_accuracy: 0.4893 - val_Rating_mae: 0.1507 - val_Year_mae: 0.1270 Epoch 25/30 147/151 [============================>.] - ETA: 0s - loss: 1.7327 - Genre_loss: 1.3941 - Rating_loss: 0.1288 - Year_loss: 0.2097 - Genre_accuracy: 0.4610 - Rating_mae: 0.2811 - Year_mae: 0.3603 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 141ms/step - loss: 1.7317 - Genre_loss: 1.3937 - Rating_loss: 0.1286 - Year_loss: 0.2094 - Genre_accuracy: 0.4612 - Rating_mae: 0.2809 - Year_mae: 0.3599 - val_loss: 1.4166 - val_Genre_loss: 1.3676 - val_Rating_loss: 0.0315 - val_Year_loss: 0.0175 - val_Genre_accuracy: 0.4834 - val_Rating_mae: 0.1507 - val_Year_mae: 0.1191 Epoch 26/30 148/151 [============================>.] - ETA: 0s - loss: 1.6570 - Genre_loss: 1.3687 - Rating_loss: 0.1089 - Year_loss: 0.1795 - Genre_accuracy: 0.4694 - Rating_mae: 0.2621 - Year_mae: 0.3336 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 140ms/step - loss: 1.6563 - Genre_loss: 1.3681 - Rating_loss: 0.1088 - Year_loss: 0.1794 - Genre_accuracy: 0.4697 - Rating_mae: 0.2620 - Year_mae: 0.3335 - val_loss: 1.4231 - val_Genre_loss: 1.3759 - val_Rating_loss: 0.0299 - val_Year_loss: 0.0172 - val_Genre_accuracy: 0.4873 - val_Rating_mae: 0.1468 - val_Year_mae: 0.1185 Epoch 27/30 149/151 [============================>.] - ETA: 0s - loss: 1.5997 - Genre_loss: 1.3298 - Rating_loss: 0.1010 - Year_loss: 0.1690 - Genre_accuracy: 0.4868 - Rating_mae: 0.2516 - Year_mae: 0.3265 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 21s 140ms/step - loss: 1.5995 - Genre_loss: 1.3297 - Rating_loss: 0.1009 - Year_loss: 0.1689 - Genre_accuracy: 0.4869 - Rating_mae: 0.2516 - Year_mae: 0.3263 - val_loss: 1.4179 - val_Genre_loss: 1.3735 - val_Rating_loss: 0.0283 - val_Year_loss: 0.0161 - val_Genre_accuracy: 0.4883 - val_Rating_mae: 0.1428 - val_Year_mae: 0.1146 Epoch 28/30 150/151 [============================>.] - ETA: 0s - loss: 1.5413 - Genre_loss: 1.3019 - Rating_loss: 0.0909 - Year_loss: 0.1484 - Genre_accuracy: 0.4963 - Rating_mae: 0.2382 - Year_mae: 0.3007 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 148ms/step - loss: 1.5413 - Genre_loss: 1.3020 - Rating_loss: 0.0909 - Year_loss: 0.1484 - Genre_accuracy: 0.4963 - Rating_mae: 0.2382 - Year_mae: 0.3006 - val_loss: 1.4187 - val_Genre_loss: 1.3755 - val_Rating_loss: 0.0282 - val_Year_loss: 0.0150 - val_Genre_accuracy: 0.4863 - val_Rating_mae: 0.1429 - val_Year_mae: 0.1109 Epoch 29/30 150/151 [============================>.] - ETA: 0s - loss: 1.4910 - Genre_loss: 1.2734 - Rating_loss: 0.0845 - Year_loss: 0.1332 - Genre_accuracy: 0.5016 - Rating_mae: 0.2297 - Year_mae: 0.2885 * * * * * Image could not be loaded - skipping * * * * * 151/151 [==============================] - 22s 148ms/step - loss: 1.4909 - Genre_loss: 1.2733 - Rating_loss: 0.0845 - Year_loss: 0.1331 - Genre_accuracy: 0.5017 - Rating_mae: 0.2297 - Year_mae: 0.2885 - val_loss: 1.4228 - val_Genre_loss: 1.3830 - val_Rating_loss: 0.0265 - val_Year_loss: 0.0133 - val_Genre_accuracy: 0.4922 - val_Rating_mae: 0.1381 - val_Year_mae: 0.1041 Epoch 30/30 151/151 [==============================] - 22s 149ms/step - loss: 1.4430 - Genre_loss: 1.2429 - Rating_loss: 0.0736 - Year_loss: 0.1265 - Genre_accuracy: 0.5225 - Rating_mae: 0.2150 - Year_mae: 0.2795 - val_loss: 1.4244 - val_Genre_loss: 1.3882 - val_Rating_loss: 0.0246 - val_Year_loss: 0.0116 - val_Genre_accuracy: 0.4873 - val_Rating_mae: 0.1325 - val_Year_mae: 0.0971 CPU times: user 13min 41s, sys: 1min 39s, total: 15min 20s Wall time: 10min 56s
content_model.model.save("Model_sigmoid")
INFO:tensorflow:Assets written to: Model_sigmoid/assets
# %tensorboard --logdir logs
from matplotlib import pyplot as plt
plt.figure(figsize=(8, 6), dpi=120)
plt.plot(history.history['Genre_loss'])
plt.plot(history.history['val_Genre_loss'])
plt.title('Model loss/Epoch - Genre')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['training', 'validation'])
plt.show()
plt.figure(figsize=(8, 6), dpi=120)
plt.plot(history.history['Genre_accuracy'])
plt.plot(history.history['val_Genre_accuracy'])
plt.title('Model accuracy/Epoch - Genre')
plt.ylabel('Accuracy')
plt.xlabel('Epoch')
plt.legend(['training', 'validation'])
plt.show()
loss_tr = history.history['Genre_loss'][-1]
accuracy_tr = history.history['Genre_accuracy'][-1]
print(f'Model train loss: {loss_tr:.4}')
print(f'Model train accuracy: {accuracy_tr:.4}')
Model train loss: 1.245 Model train accuracy: 0.5242
import numpy as np
t_gen = data_generator.generate_images(test_idx, cfgs.HYPERPARAMS.BATCH_SIZE_TEST, False)
genre_inf, rating_inf, year_inf = content_model.model.predict(t_gen,
steps=len(test_idx)//cfgs.HYPERPARAMS.BATCH_SIZE_TEST)
genres_pred = genre_inf.argmax(axis=-1)
ratings_pred = rating_inf * data_generator.max_rating
years_pred = year_inf * data_generator.max_year
genres_true, ratings_true, years_true = data_generator.generate_truth(test_idx, cfgs.HYPERPARAMS.BATCH_SIZE_TEST)
from sklearn.metrics import classification_report
cr_genre = classification_report(genres_true, genres_pred, target_names=dataset_dict['Genre_alias'].keys())
print(cr_genre)
precision recall f1-score support
Comedy 0.51 0.49 0.50 200
Action 0.44 0.39 0.41 195
Documentary 0.38 0.59 0.46 196
Crime 0.41 0.35 0.38 178
Animation 0.81 0.58 0.68 189
Horror 0.49 0.46 0.47 194
accuracy 0.48 1152
macro avg 0.50 0.48 0.48 1152
weighted avg 0.50 0.48 0.48 1152
from sklearn.metrics import r2_score
print('R2 score for ratings: ', r2_score(ratings_true, ratings_pred))
R2 score for ratings: -0.41678166774098924